Anonymity Effects: A Large-Scale Dataset from an Anonymous Social Media Platform

Abstract

Today online social media sites function as the medium of expression for billions of users. As a result, aside from conventional social media sites like Facebook and Twitter, platform designers introduced many alternative social media platforms (e.g., 4chan, Whisper, Snapchat, Mastodon) to serve specific userbases. Among these platforms, anonymous social media sites like Whisper and 4chan hold a special place for researchers. Unlike conventional social media sites, posts on anonymous social media sites are not associated with persistent user identities or profiles. Thus, these anonymous social media sites can provide an extremely interesting data-driven lens into the effects of anonymity on online user behavior. However, to the best of our knowledge, currently there are no publicly available datasets to facilitate research efforts on these anonymity effects. To that end, in this paper, we aim to publicly release the first ever large-scale dataset from Whisper, a large anonymous online social media platform. Specifically, our dataset contains 89.8 Million Whisper posts (called "whispers'') published between a 2-year period from June 6, 2014 to June 6, 2016 (when Whisper was quite popular). Each of these whispers contained both post text and associated metadata. The metadata contains information like coarse-grained location of upload and categories of whispers. We also present preliminary descriptive statistics to demonstrate a significant language and categorical diversity in our dataset. We leverage previous work as well as novel analysis to demonstrate that the whispers contain personal emotions and opinions (likely facilitated by a disinhibition complex due to anonymity). Consequently, we envision that our dataset will facilitate novel research ranging from understanding online aggression to detect depression within online populace.